A robot localization plays an important role in the field of robot navigation. One of the most commonly used localization algorithms is Monte Carlo algorithm. To improve the efficiency of robot localization, many modified algorithms have been proposed, such as Self-Adaptive Monte Carlo algorithm. However, this method requires a lot of storage space and intensive computing, especially in large environments. In recent years, because of the rapid development of cloud computing, the data can be dynamically allocated. Therefore, this paper combines the Self-Adaptive Monte Carlo Localization algorithm with cloud computing. Some experimental results illustrate the proposed architecture, which can quickly establish the map database and provide the shared map information to multiple robots. In addition, the proposed method reduces the computational load and expands the scope of activities.